An official website of the United States government
Battery-free and intermittently powered devices offer long lifetimes and enable deployment in new applications and environments. Unfortunately, developing sophisticated inference-capable applications is still challenging due to the lack of platform support for more advanced (32-bit) microprocessors and specialized accelerators---which can execute data-intensive machine learning tasks, but add complexity across the stack when dealing with intermittent power. We present Protean to bridge the platform gap for inference-capable battery-free sensors. Designed for runtime scalability, meeting the dynamic range of energy harvesters with matching heterogeneous processing elements like neural network accelerators. We develop a modular "plug-and-play" hardware platform, SuperSensor, with a reconfigurable energy storage circuit that powers a 32-bit ARM-based microcontroller with a convolutional neural network accelerator. An adaptive task-based runtime system, Chameleon, provides intermittency-proof execution of machine learning tasks across heterogeneous processing elements. The runtime automatically scales and dispatches these tasks based on incoming energy, current state, and programmer annotations. A code generator, Metamorph, automates conversion of ML models to intermittent safe execution across heterogeneous compute elements. We evaluate Protean with audio and image workloads and demonstrate up to 666x improvement in inference energy efficiency by enabling usage of modern computational elements within intermittent computing. Further, Protean provides up to 166% higher throughput compared to non-adaptive baselines.
more »
« less
Protean: Adaptive Hardware-Accelerated Intermittent Computing
Today's smart devices have short battery lifetimes, high installation and maintenance costs, and rapid obsolescence - all leading to the explosion of electronic waste in the past two decades. These problems will worsen as the number of connected devices grows to one trillion by 2035. Energy harvesting, battery-free devices offer an alternative. Getting rid of the battery reduces e-waste, promises long lifetimes, and enables deployment in new applications and environments. Unfortunately, developing sophisticated inference-capable applications is still challenging. The lack of platform support for advanced (32-bit) microprocessors and specialized accelerators, which can execute dataintensive machine-learning tasks, has held back batteryless devices.
more »
« less
- PAR ID:
- 10480082
- Editor(s):
- Ko, Steve
- Publisher / Repository:
- Association of Computing Machinery
- Date Published:
- Journal Name:
- GetMobile: Mobile Computing and Communications
- Volume:
- 27
- Issue:
- 1
- ISSN:
- 2375-0529
- Page Range / eLocation ID:
- 5 to 10
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Abstract The current change in battery technology followed by the almost immediate adoption of lithium as a key resource powering our energy needs in various applications is undeniable. Lithium-ion batteries (LIBs) are at the forefront of the industry and offer excellent performance. The application of LIBs is expected to continue to increase. The adoption of renewable energies has spurred this LIB proliferation and resulted in a dramatic increase in LIB waste. In this review, we address waste LIB collection and segregation approaches, waste LIB treatment approaches, and related economics. We have coined a “green score” concept based on a review of several quantitative analyses from the literature to compare the three mainstream recycling processes: pyrometallurgical, hydrometallurgical, and direct recycling. In addition, we analyze the current trends in policymaking and in government incentive development directed toward promoting LIB waste recycling. Future LIB recycling perspectives are analyzed, and opportunities and threats to LIB recycling are presented.more » « less
-
A digital non-Foster radio approach is proposed to mitigate Wheeler-Chu limits of electrically-small antennas, with significant potential to significantly reduce energy consumption in the VHF (very high frequency) band, where radio propagation losses below 200 MHz are 100 times less than losses above 2 GHz. Operation at lower frequency could greatly extend lifetimes of small low-power Internet-of-Things devices such as battery-powered sensors operating primarily as transmitters. Unfortunately, physical size constraints and the Wheeler-Chu limit have greatly hindered utilization of VHF bands for mobile devices, where even a 200 MHz half-wave dipole is an unwieldy 0.75 m. However, recent advances in non-Foster impedance matching methods have overcome these limits. In addition, recent digital non-Foster methods are shown to closely resemble digital radio architectures, suggesting that these newer digital non-Foster methods can be readily adopted in new digital radio designs. Therefore, a novel digital non-Foster radio architecture is proposed, where digital non-Foster methods enable small devices in energy-efficient VHF bands while overcoming Wheeler-Chu antenna-size limits.more » « less
-
Recent advancements in semiconductor technologies have stimulated the growth of ultra-low power wearable devices. However, these devices often pose critical constraints in usability and functionality because of the on-device battery as the primary power source [1]. For example, periodic charging of wearable devices hampers the continuous monitoring of users' fitness or health conditions [2], and batteries and charging equipment have been identified as one of the most rapidly growing electronic waste streams [3]. To counteract the above-mentioned complications associated with the management of on-device batteries, wireless power transmission technologies capable of charging wearable devices in a completely unobtrusive and seamless manner have become an emerging topic of research over the past decade [4]. Researchers have instrumented daily objects or the surrounding environment with equipment that can wirelessly transfer energy from a variety of sources, such as Radio Frequency (RF) signals, laser, and electromagnetic fields [5]. However, these solutions require large and costly infrastructure and/or need to transmit a significant amount of power to support reasonable power harvesting at the wearable devices, which conflict with the vision of ubiquitously available and scalable charging support.more » « less
-
Human-caused climate degradation and the explosion of electronic waste have pushed the computing community to explore fundamental alternatives to the current battery-powered, over-provisioned ubiquitous computing devices that need constant replacement and recharging. Soil Microbial Fuel Cells (SMFCs) offer promise as a renewable energy source that is biocompatible and viable in difficult environments where traditional batteries and solar panels fall short. However, SMFC development is in its infancy, and challenges like robustness to environmental factors and low power output stymie efforts to implement real-world applications in terrestrial environments. This work details a 2-year iterative process that uncovers barriers to practical SMFC design for powering electronics, which we address through a mechanistic understanding of SMFC theory from the literature. We present nine months of deployment data gathered from four SMFC experiments exploring cell geometries, resulting in an improved SMFC that generates power across a wider soil moisture range. From these experiments, we extracted key lessons and a testing framework, assessed SMFC's field performance, contextualized improvements with emerging and existing computing systems, and demonstrated the improved SMFC powering a wireless sensor for soil moisture and touch sensing. We contribute our data, methodology, and designs to establish the foundation for a sustainable, soil-powered future.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1145/3599184.3599186
